Numerous factors have contributed to declines in populations of the federally threatened Agassiz's Desert Tortoise (Gopherus agassizii) and continue to limit recovery. In 2010, we surveyed a low-density population on a military test facility in the northwestern Mojave Desert of California, USA, to evaluate population status and identify potential factors contributing to distribution and low densities. Estimated densities of live tortoises ranged spatially from 1.2/km² to 15.1/km². Although only one death of a breeding-age tortoise was recorded for the 4-yr period prior to the survey, remains of 16 juvenile and immature tortoises were found, and most showed signs of predation by Common Ravens (Corvus corax) and mammals. Predation may have limited recruitment of young tortoises into the adult size classes. To evaluate the relative importance of different types of impacts to tortoises, we developed predictive models for spatially explicit densities of tortoise sign and live tortoises using topography (i.e., slope), predators (Common Raven, signs of mammalian predators), and anthropogenic impacts (distances from paved road and denuded areas, density of ordnance fragments) as covariates. Models suggest that densities of tortoise sign increased with slope and signs of mammalian predators and decreased with Common Ravens, while also varying based on interaction effects involving these predictors as well as distances from paved roads, denuded areas, and ordnance. Similarly, densities of live tortoises varied by interaction effects among distances to denuded areas and paved roads, density of ordnance fragments, and slope. Thus multiple factors predict the densities and distribution of this population.